app.py

import os
import gradio as gr
import json
import logging
import torch
from PIL import Image
import spaces
from diffusers import DiffusionPipeline, AutoencoderTiny, AutoencoderKL, AutoPipelineForImage2Image
from live_preview_helpers import calculate_shift, retrieve_timesteps, flux_pipe_call_that_returns_an_iterable_of_images
from diffusers.utils import load_image
from huggingface_hub import hf_hub_download, HfFileSystem, ModelCard, snapshot_download
import copy
import random
import time

# Load LoRAs from JSON file
with open('loras.json', 'r') as f:
    loras = json.load(f)

# Initialize the base model
dtype = torch.bfloat16
device = "cuda" if torch.cuda.is_available() else "cpu"
base_model = "black-forest-labs/FLUX.1-dev"

taef1 = AutoencoderTiny.from_pretrained("madebyollin/taef1", torch_dtype=dtype).to(device)
good_vae = AutoencoderKL.from_pretrained(base_model, subfolder="vae", torch_dtype=dtype).to(device)
pipe = DiffusionPipeline.from_pretrained(base_model, torch_dtype=dtype, vae=taef1).to(device)
pipe_i2i = AutoPipelineForImage2Image.from_pretrained(base_model,
                                                      vae=good_vae,
                                                      transformer=pipe.transformer,
                                                      text_encoder=pipe.text_encoder,
                                                      tokenizer=pipe.tokenizer,
                                                      text_encoder_2=pipe.text_encoder_2,
                                                      tokenizer_2=pipe.tokenizer_2,
                                                      torch_dtype=dtype
                                                     )

MAX_SEED = 2**32-1

pipe.flux_pipe_call_that_returns_an_iterable_of_images = flux_pipe_call_that_returns_an_iterable_of_images.__get__(pipe)

class calculateDuration:
    def __init__(self, activity_name=""):
        self.activity_name = activity_name

    def __enter__(self):
        self.start_time = time.time()
        return self
    
    def __exit__(self, exc_type, exc_value, traceback):
        self.end_time = time.time()
        self.elapsed_time = self.end_time - self.start_time
        if self.activity_name:
            print(f"Elapsed time for {self.activity_name}: {self.elapsed_time:.6f} seconds")
        else:
            print(f"Elapsed time: {self.elapsed_time:.6f} seconds")

def update_selection(evt: gr.SelectData, selected_indices, width, height):
    selected_index = evt.index
    selected_indices = selected_indices or []
    if selected_index in selected_indices:
        # LoRA is already selected, remove it
        selected_indices.remove(selected_index)
    else:
        if len(selected_indices) < 2:
            selected_indices.append(selected_index)
        else:
            raise gr.Error("You can select up to 2 LoRAs only.")

    # Initialize outputs
    selected_info_1 = ""
    selected_info_2 = ""
    lora_scale_1 = 0.95
    lora_scale_2 = 0.95
    lora_image_1 = None
    lora_image_2 = None
    if len(selected_indices) >= 1:
        lora1 = loras[selected_indices[0]]
        selected_info_1 = f"### LoRA 1 Selected: [{lora1['title']}](https://huggingface.co/{lora1['repo']}) ✨"
        lora_image_1 = lora1['image']
    if len(selected_indices) >= 2:
        lora2 = loras[selected_indices[1]]
        selected_info_2 = f"### LoRA 2 Selected: [{lora2['title']}](https://huggingface.co/{lora2['repo']}) ✨"
        lora_image_2 = lora2['image']

    # Update prompt placeholder based on last selected LoRA
    if selected_indices:
        last_selected_lora = loras[selected_indices[-1]]
        new_placeholder = f"Type a prompt for {last_selected_lora['title']}"
    else:
        new_placeholder = "Type a prompt after selecting a LoRA"

    return (
        gr.update(placeholder=new_placeholder),
        selected_info_1,
        selected_info_2,
        selected_indices,
        lora_scale_1,
        lora_scale_2,
        width,
        height,
        lora_image_1,
        lora_image_2,
    )

def remove_lora_1(selected_indices):
    selected_indices = selected_indices or []
    if len(selected_indices) >= 1:
        selected_indices.pop(0)
    # Update selected_info_1 and selected_info_2
    selected_info_1 = ""
    selected_info_2 = ""
    lora_scale_1 = 0.95
    lora_scale_2 = 0.95
    lora_image_1 = None
    lora_image_2 = None
    if len(selected_indices) >= 1:
        lora1 = loras[selected_indices[0]]
        selected_info_1 = f"### LoRA 1 Selected: [{lora1['title']}](https://huggingface.co/{lora1['repo']}) ✨"
        lora_image_1 = lora1['image']
    if len(selected_indices) >= 2:
        lora2 = loras[selected_indices[1]]
        selected_info_2 = f"### LoRA 2 Selected: [{lora2['title']}](https://huggingface.co/{lora2['repo']}) ✨"
        lora_image_2 = lora2['image']
    return selected_info_1, selected_info_2, selected_indices, lora_scale_1, lora_scale_2, lora_image_1, lora_image_2

def remove_lora_2(selected_indices):
    selected_indices = selected_indices or []
    if len(selected_indices) >= 2:
        selected_indices.pop(1)
    # Update selected_info_1 and selected_info_2
    selected_info_1 = ""
    selected_info_2 = ""
    lora_scale_1 = 0.95
    lora_scale_2 = 0.95
    lora_image_1 = None
    lora_image_2 = None
    if len(selected_indices) >= 1:
        lora1 = loras[selected_indices[0]]
        selected_info_1 = f"### LoRA 1 Selected: [{lora1['title']}](https://huggingface.co/{lora1['repo']}) ✨"
        lora_image_1 = lora1['image']
    if len(selected_indices) >= 2:
        lora2 = loras[selected_indices[1]]
        selected_info_2 = f"### LoRA 2 Selected: [{lora2['title']}](https://huggingface.co/{lora2['repo']}) ✨"
        lora_image_2 = lora2['image']
    return selected_info_1, selected_info_2, selected_indices, lora_scale_1, lora_scale_2, lora_image_1, lora_image_2

def randomize_loras(selected_indices):
    if len(loras) < 2:
        raise gr.Error("Not enough LoRAs to randomize.")
    selected_indices = random.sample(range(len(loras)), 2)
    lora1 = loras[selected_indices[0]]
    lora2 = loras[selected_indices[1]]
    selected_info_1 = f"### LoRA 1 Selected: [{lora1['title']}](https://huggingface.co/{lora1['repo']}) ✨"
    selected_info_2 = f"### LoRA 2 Selected: [{lora2['title']}](https://huggingface.co/{lora2['repo']}) ✨"
    lora_scale_1 = 0.95
    lora_scale_2 = 0.95
    lora_image_1 = lora1['image']
    lora_image_2 = lora2['image']
    return selected_info_1, selected_info_2, selected_indices, lora_scale_1, lora_scale_2, lora_image_1, lora_image_2

@spaces.GPU(duration=70)
def generate_image(prompt_mash, steps, seed, cfg_scale, width, height, progress):
    print("Entrou aqui!")
    pipe.to("cuda")
    generator = torch.Generator(device="cuda").manual_seed(seed)
    with calculateDuration("Generating image"):
        # Generate image
        for img in pipe.flux_pipe_call_that_returns_an_iterable_of_images(
            prompt=prompt_mash,
            num_inference_steps=steps,
            guidance_scale=cfg_scale,
            width=width,
            height=height,
            generator=generator,
            joint_attention_kwargs={"scale": 1.0},
            output_type="pil",
            good_vae=good_vae,
        ):
            yield img

@spaces.GPU(duration=70)
def generate_image_to_image(prompt_mash, image_input_path, image_strength, steps, cfg_scale, width, height, seed):
    pipe_i2i.to("cuda")
    generator = torch.Generator(device="cuda").manual_seed(seed)
    image_input = load_image(image_input_path)
    final_image = pipe_i2i(
        prompt=prompt_mash,
        image=image_input,
        strength=image_strength,
        num_inference_steps=steps,
        guidance_scale=cfg_scale,
        width=width,
        height=height,
        generator=generator,
        joint_attention_kwargs={"scale": 1.0},
        output_type="pil",
    ).images[0]
    return final_image 
    
def run_lora(prompt, image_input, image_strength, cfg_scale, steps, selected_indices, lora_scale_1, lora_scale_2, randomize_seed, seed, width, height, progress=gr.Progress(track_tqdm=True)):
    if not selected_indices:
        raise gr.Error("You must select at least one LoRA before proceeding.")

    selected_loras = [loras[idx] for idx in selected_indices]

    # Build the prompt with trigger words
    prepends = []
    appends = []
    for lora in selected_loras:
        trigger_word = lora.get('trigger_word', '')
        if trigger_word:
            if lora.get("trigger_position") == "prepend":
                prepends.append(trigger_word)
            else:
                appends.append(trigger_word)
    prompt_mash = " ".join(prepends + [prompt] + appends)
    print("Prompt Mash: ", prompt_mash)
    # Unload previous LoRA weights
    with calculateDuration("Unloading LoRA"):
        pipe.unload_lora_weights()
        pipe_i2i.unload_lora_weights()

    # Load LoRA weights with respective scales
    lora_names = []
    with calculateDuration("Loading LoRA weights"):
        for idx, lora in enumerate(selected_loras):
            lora_name = f"lora_{idx}"
            lora_names.append(lora_name)
            lora_path = lora['repo']
            scale = lora_scale_1 if idx == 0 else lora_scale_2
            if image_input is not None:
                if "weights" in lora:
                    pipe_i2i.load_lora_weights(lora_path, weight_name=lora["weights"], low_cpu_mem_usage=True, adapter_name=lora_name)
                else:
                    pipe_i2i.load_lora_weights(lora_path, low_cpu_mem_usage=True, adapter_name=lora_name)
            else:
                if "weights" in lora:
                    pipe.load_lora_weights(lora_path, weight_name=lora["weights"], low_cpu_mem_usage=True, adapter_name=lora_name)
                else:
                    pipe.load_lora_weights(lora_path, low_cpu_mem_usage=True, adapter_name=lora_name)
        print(lora_names)
        if image_input is not None:
            pipe_i2i.set_adapters(lora_names, adapter_weights=[lora_scale_1, lora_scale_2])
        else:
            pipe.set_adapters(lora_names, adapter_weights=[lora_scale_1, lora_scale_2])
    # Set random seed for reproducibility
    with calculateDuration("Randomizing seed"):
        if randomize_seed:
            seed = random.randint(0, MAX_SEED)

    # Generate image
    if image_input is not None:
        final_image = generate_image_to_image(prompt_mash, image_input, image_strength, steps, cfg_scale, width, height, seed)
        yield final_image, seed, gr.update(visible=False)
    else:
        image_generator = generate_image(prompt_mash, steps, seed, cfg_scale, width, height, progress)
        # Consume the generator to get the final image
        final_image = None
        step_counter = 0
        for image in image_generator:
            step_counter+=1
            final_image = image
            progress_bar = f'<div class="progress-container"><div class="progress-bar" style="--current: {step_counter}; --total: {steps};"></div></div>'
            yield image, seed, gr.update(value=progress_bar, visible=True)
        yield final_image, seed, gr.update(value=progress_bar, visible=False)

def get_huggingface_safetensors(link):
    split_link = link.split("/")
    if len(split_link) == 2:
        model_card = ModelCard.load(link)
        base_model = model_card.data.get("base_model")
        print(base_model)
        if base_model not in ["black-forest-labs/FLUX.1-dev", "black-forest-labs/FLUX.1-schnell"]:
            raise Exception("Not a FLUX LoRA!")
        image_path = model_card.data.get("widget", [{}])[0].get("output", {}).get("url", None)
        trigger_word = model_card.data.get("instance_prompt", "")
        image_url = f"https://huggingface.co/{link}/resolve/main/{image_path}" if image_path else None
        fs = HfFileSystem()
        safetensors_name = None
        try:
            list_of_files = fs.ls(link, detail=False)
            for file in list_of_files:
                if file.endswith(".safetensors"):
                    safetensors_name = file.split("/")[-1]
                if not image_url and file.lower().endswith((".jpg", ".jpeg", ".png", ".webp")):
                    image_elements = file.split("/")
                    image_url = f"https://huggingface.co/{link}/resolve/main/{image_elements[-1]}"
        except Exception as e:
            print(e)
            raise Exception("Invalid Hugging Face repository with a *.safetensors LoRA")
        if not safetensors_name:
            raise Exception("No *.safetensors file found in the repository")
        return split_link[1], link, safetensors_name, trigger_word, image_url

def check_custom_model(link):
    if link.startswith("https://"):
        if link.startswith("https://huggingface.co") or link.startswith("https://www.huggingface.co"):
            link_split = link.split("huggingface.co/")
            return get_huggingface_safetensors(link_split[1])
    else: 
        return get_huggingface_safetensors(link)

def add_custom_lora(custom_lora, selected_indices):
    global loras
    if custom_lora:
        try:
            title, repo, path, trigger_word, image = check_custom_model(custom_lora)
            print(f"Loaded custom LoRA: {repo}")
            card = f'''
            <div class="custom_lora_card">
              <span>Loaded custom LoRA:</span>
              <div class="card_internal">
                <img src="{image}" />
                <div>
                    <h3>{title}</h3>
                    <small>{"Using: <code><b>"+trigger_word+"</code></b> as the trigger word" if trigger_word else "No trigger word found. If there's a trigger word, include it in your prompt"}<br></small>
                </div>
              </div>
            </div>
            '''
            existing_item_index = next((index for (index, item) in enumerate(loras) if item['repo'] == repo), None)
            if existing_item_index is None:
                new_item = {
                    "image": image,
                    "title": title,
                    "repo": repo,
                    "weights": path,
                    "trigger_word": trigger_word
                }
                print(new_item)
                existing_item_index = len(loras)
                loras.append(new_item)

            # Update gallery
            gallery_items = [(item["image"], item["title"]) for item in loras]
            # Update selected_indices if there's room
            if len(selected_indices) < 2:
                selected_indices.append(existing_item_index)
                selected_info_1 = ""
                selected_info_2 = ""
                lora_scale_1 = 0.95
                lora_scale_2 = 0.95
                lora_image_1 = None
                lora_image_2 = None
                if len(selected_indices) >= 1:
                    lora1 = loras[selected_indices[0]]
                    selected_info_1 = f"### LoRA 1 Selected: [{lora1['title']}](https://huggingface.co/{lora1['repo']}) ✨"
                    lora_image_1 = lora1['image']
                if len(selected_indices) >= 2:
                    lora2 = loras[selected_indices[1]]
                    selected_info_2 = f"### LoRA 2 Selected: [{lora2['title']}](https://huggingface.co/{lora2['repo']}) ✨"
                    lora_image_2 = lora2['image']
                return (gr.update(visible=True, value=card), gr.update(visible=True), gr.update(value=gallery_items),
                        selected_info_1, selected_info_2, selected_indices, lora_scale_1, lora_scale_2, lora_image_1, lora_image_2)
            else:
                return (gr.update(visible=True, value=card), gr.update(visible=True), gr.update(value=gallery_items),
                        gr.NoChange(), gr.NoChange(), selected_indices, gr.NoChange(), gr.NoChange(), gr.NoChange(), gr.NoChange())
        except Exception as e:
            print(e)
            return gr.update(visible=True, value=str(e)), gr.update(visible=True), gr.NoChange(), gr.NoChange(), gr.NoChange(), selected_indices, gr.NoChange(), gr.NoChange(), gr.NoChange(), gr.NoChange()
    else:
        return gr.update(visible=False), gr.update(visible=False), gr.NoChange(), gr.NoChange(), gr.NoChange(), selected_indices, gr.NoChange(), gr.NoChange(), gr.NoChange(), gr.NoChange()

def remove_custom_lora(custom_lora_info, custom_lora_button, selected_indices):
    global loras
    if loras:
        custom_lora_repo = loras[-1]['repo']
        # Remove from loras list
        loras = loras[:-1]
        # Remove from selected_indices if selected
        custom_lora_index = len(loras)
        if custom_lora_index in selected_indices:
            selected_indices.remove(custom_lora_index)
    # Update gallery
    gallery_items = [(item["image"], item["title"]) for item in loras]
    # Update selected_info and images
    selected_info_1 = ""
    selected_info_2 = ""
    lora_scale_1 = 0.95
    lora_scale_2 = 0.95
    lora_image_1 = None
    lora_image_2 = None
    if len(selected_indices) >= 1:
        lora1 = loras[selected_indices[0]]
        selected_info_1 = f"### LoRA 1 Selected: [{lora1['title']}](https://huggingface.co/{lora1['repo']}) ✨"
        lora_image_1 = lora1['image']
    if len(selected_indices) >= 2:
        lora2 = loras[selected_indices[1]]
        selected_info_2 = f"### LoRA 2 Selected: [{lora2['title']}](https://huggingface.co/{lora2['repo']}) ✨"
        lora_image_2 = lora2['image']
    return gr.update(visible=False), gr.update(visible=False), gr.update(value=gallery_items), selected_info_1, selected_info_2, selected_indices, lora_scale_1, lora_scale_2, lora_image_1, lora_image_2

run_lora.zerogpu = True

css = '''
#gen_btn{height: 100%}
#title{text-align: center}
#title h1{font-size: 3em; display:inline-flex; align-items:center}
#title img{width: 100px; margin-right: 0.5em}
#gallery .grid-wrap{height: 10vh}
#lora_list{background: var(--block-background-fill);padding: 0 1em .3em; font-size: 90%}
.custom_lora_card{margin-bottom: 1em}
.card_internal{display: flex;height: 100px;margin-top: .5em}
.card_internal img{margin-right: 1em}
.styler{--form-gap-width: 0px !important}
#progress{height:30px}
#progress .generating{display:none}
.progress-container {width: 100%;height: 30px;background-color: #f0f0f0;border-radius: 15px;overflow: hidden;margin-bottom: 20px}
.progress-bar {height: 100%;background-color: #4f46e5;width: calc(var(--current) / var(--total) * 100%);transition: width 0.5s ease-in-out}
.button_total{height: 100%}
#loaded_loras [data-testid="block-info"]{font-size:80%}
'''

with gr.Blocks(theme=gr.themes.Soft(), css=css, delete_cache=(60, 3600)) as app:
    title = gr.HTML(
        """<h1><img src="https://huggingface.co/spaces/multimodalart/flux-lora-the-explorer/resolve/main/flux_lora.png" alt="LoRA"> LoRA Lab</h1>""",
        elem_id="title",
    )
    selected_indices = gr.State([])
    with gr.Row():
        with gr.Column(scale=3):
            prompt = gr.Textbox(label="Prompt", lines=1, placeholder="Type a prompt after selecting a LoRA")
        with gr.Column(scale=1):
            generate_button = gr.Button("Generate", variant="primary", elem_classes=["button_total"])
    with gr.Row(elem_id="loaded_loras"):
        with gr.Column(scale=1, min_width=25):
            randomize_button = gr.Button("🎲", variant="secondary", scale=1)
        with gr.Column(scale=8):
            with gr.Row():
                with gr.Column(scale=0, min_width=50):
                    lora_image_1 = gr.Image(label="LoRA 1 Image", interactive=False, min_width=50, width=50, show_label=False, show_share_button=False, show_download_button=False, show_fullscreen_button=False, height=50)
                with gr.Column(scale=3, min_width=100):
                    selected_info_1 = gr.Markdown("Select a LoRA 1")
                with gr.Column(scale=5, min_width=50):
                    lora_scale_1 = gr.Slider(label="LoRA 1 Scale", minimum=0, maximum=3, step=0.01, value=0.95)
            with gr.Row():
                remove_button_1 = gr.Button("Remove", size="sm")
        with gr.Column(scale=8):
            with gr.Row():
                with gr.Column(scale=0, min_width=50):
                    lora_image_2 = gr.Image(label="LoRA 2 Image", interactive=False, min_width=50, width=50, show_label=False, show_share_button=False, show_download_button=False, show_fullscreen_button=False, height=50)
                with gr.Column(scale=3, min_width=100):
                    selected_info_2 = gr.Markdown("Select a LoRA 2")
                with gr.Column(scale=2, min_width=50):
                    lora_scale_2 = gr.Slider(label="LoRA 2 Scale", minimum=0, maximum=3, step=0.01, value=0.95)
            with gr.Row():
                remove_button_2 = gr.Button("Remove", size="sm")
    with gr.Row():
        with gr.Column():
            gallery = gr.Gallery(
                [(item["image"], item["title"]) for item in loras],
                label="LoRA Gallery",
                allow_preview=False,
                columns=3,
                elem_id="gallery"
            )
            with gr.Group():
                custom_lora = gr.Textbox(label="Custom LoRA", info="LoRA Hugging Face path", placeholder="multimodalart/vintage-ads-flux")
                gr.Markdown("[Check the list of FLUX LoRAs](https://huggingface.co/models?other=base_model:adapter:black-forest-labs/FLUX.1-dev)", elem_id="lora_list")
            custom_lora_info = gr.HTML(visible=False)
            custom_lora_button = gr.Button("Remove custom LoRA", visible=False)
        with gr.Column():
            progress_bar = gr.Markdown(elem_id="progress", visible=False)
            result = gr.Image(label="Generated Image")
    with gr.Row():
        with gr.Accordion("Advanced Settings", open=False):
            with gr.Row():
                input_image = gr.Image(label="Input image", type="filepath")
                image_strength = gr.Slider(label="Denoise Strength", info="Lower means more image influence", minimum=0.1, maximum=1.0, step=0.01, value=0.75)
            with gr.Column():
                with gr.Row():
                    cfg_scale = gr.Slider(label="CFG Scale", minimum=1, maximum=20, step=0.5, value=3.5)
                    steps = gr.Slider(label="Steps", minimum=1, maximum=50, step=1, value=28)
                
                with gr.Row():
                    width = gr.Slider(label="Width", minimum=256, maximum=1536, step=64, value=1024)
                    height = gr.Slider(label="Height", minimum=256, maximum=1536, step=64, value=1024)
                
                with gr.Row():
                    randomize_seed = gr.Checkbox(True, label="Randomize seed")
                    seed = gr.Slider(label="Seed", minimum=0, maximum=MAX_SEED, step=1, value=0, randomize=True)
    
    gallery.select(
        update_selection,
        inputs=[selected_indices, width, height],
        outputs=[prompt, selected_info_1, selected_info_2, selected_indices, lora_scale_1, lora_scale_2, width, height, lora_image_1, lora_image_2]
    )
    remove_button_1.click(
        remove_lora_1,
        inputs=[selected_indices],
        outputs=[selected_info_1, selected_info_2, selected_indices, lora_scale_1, lora_scale_2, lora_image_1, lora_image_2]
    )
    remove_button_2.click(
        remove_lora_2,
        inputs=[selected_indices],
        outputs=[selected_info_1, selected_info_2, selected_indices, lora_scale_1, lora_scale_2, lora_image_1, lora_image_2]
    )
    randomize_button.click(
        randomize_loras,
        inputs=[selected_indices],
        outputs=[selected_info_1, selected_info_2, selected_indices, lora_scale_1, lora_scale_2, lora_image_1, lora_image_2]
    )
    custom_lora.change(
        add_custom_lora,
        inputs=[custom_lora, selected_indices],
        outputs=[custom_lora_info, custom_lora_button, gallery, selected_info_1, selected_info_2, selected_indices, lora_scale_1, lora_scale_2, lora_image_1, lora_image_2]
    )
    custom_lora_button.click(
        remove_custom_lora,
        inputs=[custom_lora_info, custom_lora_button, selected_indices],
        outputs=[custom_lora_info, custom_lora_button, gallery, selected_info_1, selected_info_2, selected_indices, lora_scale_1, lora_scale_2, lora_image_1, lora_image_2]
    )
    gr.on(
        triggers=[generate_button.click, prompt.submit],
        fn=run_lora,
        inputs=[prompt, input_image, image_strength, cfg_scale, steps, selected_indices, lora_scale_1, lora_scale_2, randomize_seed, seed, width, height],
        outputs=[result, seed, progress_bar]
    )

app.queue()
app.launch()